aboutsummaryrefslogtreecommitdiff
path: root/sh
Commit message (Collapse)AuthorAgeFilesLines
* Switch back to separate assignment/exportTom Ryder2022-08-141-9/+15
|
* Better way to elide hostname prompt in screen/tmuxTom Ryder2022-08-091-4/+3
|
* Use long GDB optionTom Ryder2022-07-301-1/+1
|
* Tidy up comments for POSIX interactive scriptsTom Ryder2022-07-293-0/+4
|
* Move x() shortcut to X11 directory and targetTom Ryder2022-07-291-7/+0
|
* Remove misleading commentTom Ryder2022-07-291-2/+0
|
* Add language and locale settings to profileTom Ryder2022-07-291-1/+3
|
* Massively cut down profile scriptsTom Ryder2022-07-2910-160/+23
| | | | Let's simplify all this a bit.
* Remove disused $OS env var and switchTom Ryder2022-07-292-13/+0
|
* Add local bindir even if it doesn't exist yetTom Ryder2022-07-271-4/+2
| | | | | Doesn't do any harm, as the directory is silently skipped if it doesn't exist.
* Bump up shell history limitsTom Ryder2022-05-271-2/+2
| | | | Stuff I want keeps disappearing off the buffer on more active computers.
* Lean on Debian alternatives system a bit moreTom Ryder2022-05-061-3/+9
| | | | | I'll refactor this if I ever need to (i.e. if I end up running X on something other than Debian a lot).
* Switch .xinitrc to .xsessionTom Ryder2022-04-271-2/+2
| | | | | | | | | | | | | | | | | | This is Debian-specific, but that's the only system with which I'm presently using X11 anyway. From Debian's startx(1): > Note that in the Debian system, what many people traditionally put in > the .xinitrc file should go in .xsession instead; this permits the > same X environment to be presented whether startx, xdm, or xinit is > used to start the X session. All discussion of the .xinitrc file in > the xinit(1) manual page applies equally well to .xsession. Keep in > mind that .xinitrc is used only by xinit(1) and completely ignored by > xdm(1). Indeed, everything just seems to work a bit better, probably because problems are more likely to be able to find the user instance of dbus.
* Add Torsocks management function `tor`Tom Ryder2022-03-261-0/+33
|
* Shorten some long lines in path.shTom Ryder2022-03-261-3/+5
|
* Add --no-warnings to rlwrap callsTom Ryder2021-12-241-1/+1
|
* Use inline integers rather than bit shiftingTom Ryder2020-11-131-1/+1
| | | | Syntax highlighting doesn't deal with it well
* Use better location for verse cacheTom Ryder2020-06-231-5/+5
|
* Use blank value for default fortunesTom Ryder2020-06-231-1/+1
|
* Simplify welcome.sh feature testTom Ryder2020-06-231-2/+1
|
* Apply XDG basedirs to welcome.shTom Ryder2020-06-231-7/+12
|
* Use XDG path for login welcome configTom Ryder2020-06-101-1/+1
|
* Adjust ~/.hushlogin test in login welcome scriptTom Ryder2020-06-101-2/+2
|
* Move systemd editor setting to after visualTom Ryder2020-05-122-4/+6
| | | | Loaded out of order otherwise, so doesn't get set.
* Try a compatibility switch on vi/vimTom Ryder2020-05-121-2/+7
|
* Force unwanted --quoting-style option to ls(1) offTom Ryder2020-05-022-0/+5
|
* Correct misplaced `exit` call in .profile.d fileTom Ryder2020-04-281-1/+1
| | | | | This was closing my connection immediately when I logged in interactively to any machine without systemd!
* Set systemctl editor to vi rather than edTom Ryder2020-04-271-0/+4
|
* Prevent me from using `which`Tom Ryder2019-06-101-0/+4
|
* Fix up a few cosmetic shell anti-patternsTom Ryder2019-03-291-1/+1
|
* Return to vi as default visual editorTom Ryder2019-01-071-7/+2
| | | | | | This is getting in the way of my work a bit too much. I'm still learning how to use it, but throwing myself into the deep end at this point turned out to be a bad idea.
* Further attempts at sane Emacs VISUAL settingTom Ryder2019-01-041-21/+5
|
* Block stderr from emacs daemon startupTom Ryder2019-01-021-1/+1
|
* Use short switches for emacs server pgrepTom Ryder2019-01-021-2/+1
| | | | | Turns out the pgrep from procps version 3.2.8 doesn't support these longer options.
* Add emacsclient logic to $VISUAL setTom Ryder2019-01-021-4/+20
|
* Switch to using GNU Emacs on development machinesTom Ryder2019-01-011-2/+8
| | | | | | | | | | | | | | > Andromeda, by Perseus sav'd and wed, > Hanker'd each day to see the Gorgon's head: > Till o'er a fount he held it, bade her lean, > And mirror'd in the wave was safely seen > That death she liv'd by. > Let not thine eyes know > Any forbidden thing itself, although > It once should save as well as kill: but be > Its shadow upon life enough for thee. > > --Dante Gabriel Rossetti
* Add clarifying commentTom Ryder2018-12-311-1/+1
|
* Translate a short-circuit into a conditionalTom Ryder2018-12-311-1/+3
| | | | | I only want to use the `||` short-circuits for control flow changes (return, continue, break etc).
* Add a cheeky error message to sd()Tom Ryder2018-12-311-0/+6
|
* Strip trailing slashes from sd() targetTom Ryder2018-12-311-0/+8
|
* Correct error message from sd()Tom Ryder2018-12-311-1/+1
|
* Two-space sentences in shell commentsTom Ryder2018-12-278-10/+10
|
* Follow ShellCheck's recommendation of -z over !-nTom Ryder2018-12-172-2/+2
| | | | I think !-n is a little clearer, but -z is OK.
* Adjust PS1-PS4 resetTom Ryder2018-12-171-5/+12
|
* Use exec for clearer subshell semanticsTom Ryder2018-12-171-1/+1
|
* Use `command` consistently, silence ShellCheckTom Ryder2018-12-1711-13/+23
| | | | | | | | | | ShellCheck (SC2164) is upset about these `cd` commands where the return type isn't being checked, but they're all by design, as they're the last command in the function, and thereby constitute the function's return value implicitly. Otherwise, this commit changes the shrc.d and profile.d subfiles to use the `command` wrapper only where it's actually needed.
* Refactor some conditionalsTom Ryder2018-12-1115-48/+91
|
* Factor out zsh ENV hack into one fileTom Ryder2018-12-111-8/+0
| | | | | It's a bit silly to have this in ~/.profile; it doesn't need to be there for such a niche case.
* Refactor "path list" not to require a subshellTom Ryder2018-12-111-8/+9
|
* Remove unneeded semicolon from sh "for VAR ; do"Tom Ryder2018-12-061-1/+1
| | | | | It turns out the semicolon belongs to the "in" syntax, and is optional without it.